Apply a mass change to existing Salesforce records (status normalization, owner rebalance, field cleanup) through Data Loader with validation rules and triggers respected.
- Pull a baseline export
Run a Data Loader Export of the records you intend to update. Include Id and every field you plan to change. Save the CSV as a rollback reference.
- Build the update CSV
Create a CSV with the Id column and one column per field you are changing. Leave other fields out; the update only writes the columns you include.
- Test in a sandbox
Run the same update against the same record IDs in a sandbox first. Confirm validation rules pass, triggers behave, and the updated count matches your expectation.
- Update production with batch limits
Set the Data Loader batch size to a value your triggers can handle (200 is safe; 50 for heavy logic). Run the Update operation against production. Save the success and error files.
- Investigate any errors
The error file lists the row, the field, and the error message. Fix the source data or the validation rule and re-run only the failed rows.
- Audit the result
Re-run the original export query, compare against the update CSV, and confirm field history or audit trail captured each change. Archive both files for compliance.
Data Loader setting (1 to 200 in standard mode). Smaller batches are safer for heavy triggers; larger are faster.
Switches Data Loader to the Bulk API. Faster for large jobs but asynchronous and harder to diagnose row-by-row.
Option that controls whether blank cells in the CSV null out fields or skip them.
Update modifies existing records only; Upsert can also insert new ones based on an External ID.
- Insert Null Values defaults to off. A blank cell in the update CSV is skipped unless this flag is enabled, which can leave stale data on records you thought you were clearing.
- Triggers and validation rules fire on every batch. Heavy logic on 200-record batches can exceed governor limits; size batches to your trigger profile.
- Field history tracks updates only for fields explicitly marked. If you need audit evidence for a field, enable tracking before the bulk update, not after.
- Updates that change OwnerId or sharing-critical fields can trigger large sharing recalculations. Enable defer sharing calculations for big jobs to avoid timing out the user session.